Feedback and support: TensorBlock's Twitter/X, Telegram Group and Discord server
bigscience/mt0-xxl-mt - GGUF
This repo contains GGUF format model files for bigscience/mt0-xxl-mt.
The files were quantized using machines provided by TensorBlock, and they are compatible with llama.cpp as of commit ec7f3ac.
Prompt template
Model file specification
Filename | Quant type | File Size | Description |
---|---|---|---|
mt0-xxl-mt-Q2_K.gguf | Q2_K | 5.079 GB | smallest, significant quality loss - not recommended for most purposes |
mt0-xxl-mt-Q3_K_S.gguf | Q3_K_S | 5.960 GB | very small, high quality loss |
mt0-xxl-mt-Q3_K_M.gguf | Q3_K_M | 6.397 GB | very small, high quality loss |
mt0-xxl-mt-Q3_K_L.gguf | Q3_K_L | 6.791 GB | small, substantial quality loss |
mt0-xxl-mt-Q4_0.gguf | Q4_0 | 7.540 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
mt0-xxl-mt-Q4_K_S.gguf | Q4_K_S | 7.564 GB | small, greater quality loss |
mt0-xxl-mt-Q4_K_M.gguf | Q4_K_M | 8.085 GB | medium, balanced quality - recommended |
mt0-xxl-mt-Q5_0.gguf | Q5_0 | 9.027 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
mt0-xxl-mt-Q5_K_S.gguf | Q5_K_S | 9.027 GB | large, low quality loss - recommended |
mt0-xxl-mt-Q5_K_M.gguf | Q5_K_M | 9.308 GB | large, very low quality loss - recommended |
mt0-xxl-mt-Q6_K.gguf | Q6_K | 10.607 GB | very large, extremely low quality loss |
mt0-xxl-mt-Q8_0.gguf | Q8_0 | 13.736 GB | very large, extremely low quality loss - not recommended |
Downloading instruction
Command line
Firstly, install Huggingface Client
pip install -U "huggingface_hub[cli]"
Then, downoad the individual model file the a local directory
huggingface-cli download tensorblock/mt0-xxl-mt-GGUF --include "mt0-xxl-mt-Q2_K.gguf" --local-dir MY_LOCAL_DIR
If you wanna download multiple model files with a pattern (e.g., *Q4_K*gguf
), you can try:
huggingface-cli download tensorblock/mt0-xxl-mt-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf'
- Downloads last month
- 103
Model tree for tensorblock/mt0-xxl-mt-GGUF
Base model
bigscience/mt0-xxl-mtDatasets used to train tensorblock/mt0-xxl-mt-GGUF
Evaluation results
- Accuracy on Winogrande XL (xl)validation set self-reported62.670
- Accuracy on XWinograd (en)test set self-reported83.310
- Accuracy on XWinograd (fr)test set self-reported78.310
- Accuracy on XWinograd (jp)test set self-reported80.190
- Accuracy on XWinograd (pt)test set self-reported80.990
- Accuracy on XWinograd (ru)test set self-reported79.050
- Accuracy on XWinograd (zh)test set self-reported82.340
- Accuracy on ANLI (r1)validation set self-reported49.500
- Accuracy on ANLI (r2)validation set self-reported42.000
- Accuracy on ANLI (r3)validation set self-reported48.170